Modeling Natural Killer Cell Targeted Immunotherapies
Total Page:16
File Type:pdf, Size:1020Kb
View metadata, citation and similar papers at core.ac.uk brought to you by CORE provided by Frontiers - Publisher Connector REVIEW published: 29 March 2017 doi: 10.3389/fimmu.2017.00370 Modeling Natural Killer Cell Targeted Immunotherapies Silvia Lopez-Lastra1,2,3* and James P. Di Santo1,2 1 Innate Immunity Unit, Institut Pasteur, Paris, France, 2 Inserm U1223, Paris, France, 3 Université Paris-Sud (Paris-Saclay), Paris, France Animal models have extensively contributed to our understanding of human immunobi- ology and to uncover the underlying pathological mechanisms occurring in the develop- ment of diseases. However, mouse models do not reproduce the genetic and molecular complexity inherent in human disease conditions. Human immune system (HIS) mouse models that are susceptible to human pathogens and can recapitulate human hemato- poiesis and tumor immunobiology provide one means to bridge the interspecies gap. Natural killer cells are the founding member of the innate lymphoid cell family. They exert a rapid and strong immune response against tumor and pathogen-infected cells. Their antitumor features have long been exploited for therapeutic purposes in the context of cancer. In this review, we detail the development of highly immunodeficient mouse Edited by: strains and the models currently used in cancer research. We summarize the latest Ulrike Koehl, improvements in adoptive natural killer (NK) cell therapies and the development of novel Hannover Medical School, Germany NK cell sources. Finally, we discuss the advantages of HIS mice to study the interactions Reviewed by: between human NK cells and human cancers and to develop new therapeutic strategies. Lutz Walter, Leibniz-Institute for Primate Keywords: humanized mouse models, innate lymphoid cell, natural killer cells, cancer immunotherapy, natural Research, Germany killer cell immunotherapy Philippe Saas, Etablissement Français du Sang BFC, France INTRODUCTION *Correspondence: Silvia Lopez-Lastra Since the generation of the first inbred mouse strains in the early 20th century, mice have served as [email protected] model organisms to study mammalian biology. This approach has given birth to some of the most important scientific breakthroughs and discoveries that, in many cases, led to the development of Specialty section: successful treatments for previously untreatable diseases (e.g., acute promyelocytic leukemia) (1). This article was submitted to However, Mus musculus and Homo sapiens have been evolving divergently for 85 million years, Alloimmunity and Transplantation, adapting to very different environments and undergoing selection for many traits, from the circa- a section of the journal Frontiers in Immunology dian rhythm to our body size (2). Thanks to the genome decoding, we can now appreciate that the one fifth of the genetic divergence between mice and humans is enriched in regions implicated in Received: 01 February 2017 the immune system, metabolic processes, and stress responses (3). It is, therefore, not surprising Accepted: 14 March 2017 Published: 29 March 2017 that only less than 8% of the cancer studies in animal models reach clinical trials and that more than 80% of these eventually fail when tested in humans (4). The increasing knowledge of the molecular Citation: differences between mice and humans should allow us to evaluate the degree in which animal Lopez-Lastra S and Di Santo JP (2017) Modeling Natural Killer Cell models may be suitable for translational research and when this is not the case, to then search for Targeted Immunotherapies. better systems. Front. Immunol. 8:370. With this aim, mice have been “humanized” by introducing human genes or genomic regions and doi: 10.3389/fimmu.2017.00370 by transferring human tissues or cells to study various aspects of human biology. The engraftment Frontiers in Immunology | www.frontiersin.org 1 March 2017 | Volume 8 | Article 370 Lopez-Lastra and Di Santo Modeling NK Cell Targeted Immunotherapies of human blood cells or blood-forming cells and organs into treatment. This model offers a great platform for screening of immunodeficient mice has opened a new era for translational immune adjuvants and DC targeting therapies (24). immunology and the improvement of immunotherapies against human cancer and infectious diseases caused by pathogens with HUMAN CANCER MODELS IN exclusive human tropism, such as HIV, HBV, and HCV. “HUMANIZED” MICE DEVELOPING HUMAN IMMUNE SYSTEM Immunodeficient mice allow great flexibility for the study of (HIS) MICE human tumor immunobiology. Human tumors can be generated in NSG, NOG, BRGS, and other strains using established tumor Since the discovery of the nude athymic mutations in the 1960s, cell lines, after transplantation of human primary tumors or fol- our knowledge of the host immune system and its ability to lowing de novo induction of hematological neoplasms (Figure 1). reject xenografts have led to the development of several mouse These different models provide systems that better reflect the strains that permit long-term “take” and function of the human complexity of the disease. In order to allow human tumor to tissue grafts 5( ). Experiments performed in the 1980s with severe engraft and grow in mice, the host immune system is generally combined immunodeficient (SCID) mice (that lacked functional compromised leading to tumor kinetics that may not reflect mouse adaptive lymphocytes due to mutations in the DNA- the true patient situation. As discussed earlier, human immune dependent protein kinase Prkdc) showed that these mice could components can be generated in vivo from human HSCs or other be reconstituted with human peripheral blood mononuclear cells progenitors and “supported or potentiated” later on or infused (PBMCs) or hematopoietic stem cells (HSCs) (6, 7). However, once the tumor is established. These approaches provide “mixed” some residual adaptive (leakiness) and an essentially intact innate systems in which human immune cells and human tumors can immunity in SCID mice limited the complete reconstitution of all co-exist allowing the dissection of immune deviation as well as human immune subsets. Moreover, SCID mice failed to engraft studying immunotherapy. human tumor xenografts, thereby limiting the development of A wide range of established tumor cell lines from different preclinical cancer models. An alternative system with analogous origins (brain, colon, breast, melanoma, ovarian, prostate, etc.) immunodeficiency was obtained by mutating the recombinant have been engrafted in immunocompromised mice and have activating genes (Rag1, Rag2) loci that avoided genetic “leakiness” greatly contributed to drug development and the preclinical and, in contrast to SCID mice, did not result in host radiosensitiv- assessment of potential therapies. However, the gradual accumu- ity (8, 9). Additional genetic modifications followed to further the lation of genetic and phenotypic aberrations in these cells due immunodeficiency of host mice in order to promote tolerance to to their long-term culture impacts the surface markers and the human cells. Two breakthroughs have remarkably boosted the tumorigenicity of the malignancy (25). These limitations have advancement of the field. First, Greiner and colleagues found that set aside these models to preliminary studies addressing specific the NOD strain supported an enhanced tolerance compared to questions like the ability of a potential therapy to target a certain other strains and, several years later, Takenaka’s team revealed that molecule that has been overexpressed in the cell line. In recent the molecular basis for this lies in the signal regulatory protein years, the field has been, therefore, switching toward the engraft- alpha (Sirpa) allele polymorphism (10–13). Contrarily to other ment of patient-derived primary tumors (PDX, patient-derived strains, SIRPα from NOD mice binds to human CD47 ligand xenografts) that retain the phenotypic and genetic complexity triggering a negative signal in mouse macrophages that prevents observed in clinical samples thus better predicting drug efficacy their phagocytosis (13, 14). This finding prompted the generation and clinical translatability (26, 27). These include tumor stromal of transgenic mice expressing the human or NOD strain Sirpa cells and tumor-associated lymphocytes that contribute greatly allele thus conferring enhanced human cell engraftment (15–17). to tumor growth and metastasis and, therefore, to the therapeutic The second turning point for achieving a successful xenotrans- response. These PDX-HIS mouse models can engraft the tumor as plantation was the common cytokine receptor gamma chain (γc, efficiently as the non-humanized mice, they respond to standard encoded at Il2rg), which leads to complete impairment of natural chemotherapeutic drugs similarly to patients and they have killer (NK) cell development and dendritic cell (DC) dysfunction proven to be responsive to newly derived immune modulators. (18, 19). Mice carrying Il2rg mutations were developed in various One of the better-characterized PDX models is the AML that genetic backgrounds [NSG or NOG (both NOD PrkdcSCIDIl2rg−/−) has contributed to the identification of leukemia stem cells (LSC) and BRG (Balb/c Rag2−/−Il2rg−/−)] allowing robust, long-lasting de by transplanting different stem-like cell fractions and analyzing novo multilineage development of the HIS, including human thy- the leukemia-initiating activity of each in SCID mice (28–31). mopoiesis, and are the basis for most of the currently